If the signals of a developing cancer are detected early, treatments stand the best chance of improving survival and limiting harm. As it stands, roughly half of all cancers are diagnosed at an advanced stage. Circulating extracellular vesicles (EVs) and cell-free messenger RNA (cf-mRNA) have emerged as promising biomarkers of incipient tumors. Studies are revealing their roles in cell-to-cell communication and their ability to reflect phenotypic changes from cells and tissues of origin. Seminal work by Dr. Ngo et al. demonstrated that certain cf-mRNA signatures from a simple blood draw predicted fetal age and even discriminated women at risk of preterm delivery from those who delivered at full term1. Given the power of cf-mRNA to accurately assess the progression of individual pregnancies based on tissue of origin, we hypothesized that cf-mRNA profiling may distinguish cancers from pre-malignant conditions and noncancer cohorts, enabling the early cancer detection.
To reliably identify cf-mRNA signatures, we systematically evaluated the effect of preanalytical variations on the yield and purity of EVs and cf-mRNA in human plasma2. We assessed cf-mRNA and EV profiling through differential centrifugation to deplete residual platelets and any post-freeze/thaw processing artefacts. We found that freezing plasma after depletion of residual platelets was critical, preventing the release of ex-vivo generated EV subpopulations and cf-mRNA transcripts from platelets. After investigating optimal plasma processing, we set out to understand the inter- and intra- personal baselines of these biomarkers. Through serial blood draws within and over multiple days, we demonstrated remarkable consistency in cf-mRNA signatures, with no significant variations due to diurnal cycles3. We next analyzed cf-mRNA using machine learning approaches to differentiate the presence of pcancer via liquid biopsy4. To better demonstrate the disease transitions from healthy to cancer, we included two premalignant conditions as high-risk cohorts: liver cirrhosis with liver cancer and multiple gammopathy of undetermined significance (MGUS) with multiple myeloma. We identified sets of cf-mRNA that could differentiate cancer versus healthy controls, and even healthy versus high-risk diseases. Intriguingly, we also revealed gene sets whose expression increased through specific disease progression. Through classification modeling, the accuracy of these gene sets to correctly identifying cancers was >85%, even when comparing multiple groups simultaneously.
To exploit cf-mRNA as a non-invasive biomarker, we further examined whether EVs that are known for cell-to-cell communications can also play important roles in carrying disease specific cf-mRNA. Using size-exclusion chromatography to separate EVs from RNA-binding proteins and lipoproteins, we performed transcriptomic analysis of size-fractionated plasma samples from healthy, lung cancer, liver cancer, and multiple myeloma5. Traditional RNAseq normalization requires a set of biological samples with similar amounts of RNA. However, we observed significant variations in relative cf-mRNA expression levels across plasma fractions, which were falsely amplified or regressed when using traditional global gene expression analysis. Therefore, we utilized External RNA Control Consortium (ERCC) spike-ins to control for processing variations and normalization, validating the cf-mRNA enrichment in EVs. Using immunoprecipitation followed by western blotting and qRT-PCR, we found EVs are major cf-mRNA carriers. We further validated that cf-mRNA are protected inside EVs from endogenous RNase in human plasma through RNase and detergent treatments with bare tissue RNA as controls. Finally, through assessing underlying biology of EVs and cf-mRNA in blood, we determined how selective packaging of cell free RNA inside extracellular vesicles is dysregulated in cancer. We performed differential expression analysis between each cancer type and healthy controls within specific fractions and calculated fold changes of these differentially expressed (DE) genes. Intriguingly, we found distinct enrichment patterns of cancer-differentiating cf-mRNA within specific fractions with their own biological meanings. Expanding the analyses through multiple cancer types (liver cancer, lung cancer, and multiple myeloma), we identified enrichment patterns of unique gene sets which were found in specific fractions from each cancer type.
We were astonished to discover that cf-mRNA associated with disease circulates within EVs in patient plasma. These disease-specific EVs harboring cf-mRNA can play a role in shaping the premetastatic niche to spread cancer or promote tumor growth by evading immune systems. The destination of these EVs with disease specific cf-mRNA still remains a critical, missing puzzle piece, raising questions about the reasons why cancer cells secrete such EVs. By pursuing the answers and unraveling the roles of these EVs, we strive to unlock crucial insights into cancer progression and innovative therapeutic strategies.
1 Ngo, T. T. et al. Noninvasive blood tests for fetal development predict gestational age and preterm delivery. Science (New York, N.Y.) 360, 1133-1136 (2018).
2 Kim, H. J. et al. Irreversible alteration of extracellular vesicle and cell-free messenger RNA profiles in human plasma associated with blood processing and storage. Scientific Reports 12, 2099 (2022).
3 Wagner, J. T. et al. Diurnal stability of cell-free DNA and cell-free RNA in human plasma samples. Scientific Reports 10, 16456 (2020).
4 Roskams-Hieter, B. and Kim, H. J. et al. Plasma cell-free RNA profiling distinguishes cancers from pre-malignant conditions in solid and hematologic malignancies. NPJ Precision Oncology 6, 28 (2022).
5 Kim, H. J. et al. Selective enrichment of plasma cell-free messenger RNA in cancer-associated extracellular vesicles. Commun Biol 6, 885 (2023).